Event extraction is an important research topic for information extraction. The event extraction technique extracts events of interest from the unstructured information and presents them to the user in a structured manner. Digital Library Promotion Project launched by National Library of China collects public cultural event news released by Chinese public libraries, and this paper presents the method of collecting events from cultural domain news text and propose an approach to discover cultural event automatically. To distinguish cultural news from noisy data and identify the event types, we established a cultural keywords dictionary. For event argument recognition, we adopted pattern matching method. Experimental result valid the method, and a F1 score of 86.43% on average is achieved with cultural event news from 319 Chinese public libraries.

Cultural Event Extraction for Chinese Public Libraries

AbstractWith the continuous update and development of the Internet industry, a large number of news event information of communication has emerged, and the news event information in the objective world is also increasing by hundreds of millions every day. People usually take events as a unit to carry out thinking activities. However, the increase in the amount of event information makes it difficult for people to accurately and quickly find the rules between events. Offered to solve these problems, event classes for the study of the basic unit of knowledge representation, the trigger word, class from the event object elements and environment were put forward based on synonym word event trigger word candidates found, Lin found based on big words Lin event class object elements and is based on the analysis of the interdependence syntactic event classes environmental elements found method. The experimental results show that the proposed methods for the three elements are effective. It provides a basis for theoretical research on events, and also contributes to the construction of event knowledge graph for practical application in the later stage.KeywordsEventEvent classEvent generalizationCorpus 

Event Generalization Based on Chinese Event Corpus

When adopting traditional automatic summarization, it emerged information redundancy and incomplete content covering, but currently the mainstream automatic summarization turned towards to extracting words. This paper studied the effectiveness of this issue about event elements on the size of event. Firstly obtaining the event elements through the tagged CEC corpus, then building an event elements network, calculating each node importance of the event elements network, finally getting the concise summary sentences and outputting the text summarization in accordance with the original text sequence. Experiments were conducted on CEC corpus, recall and precision had got better results to many other methods and the average value of F of this method can be raised to 0.63, which can better generalize the text content.

Tao Liao

Papers

Extraction Method of Automatic Summarization Based on Event Elements Network